Author + information
- Received June 23, 2003
- Revision received July 1, 2003
- Accepted July 22, 2003
- Published online November 19, 2003.
- ↵*Reprint requests and correspondence:
Dr. Shahbudin H. Rahimtoola, University of Southern California, 2025 Zonal Avenue, Los Angeles, California 90033, USA.
Currently, a selection of good, albeit not perfect, prosthetic heart valves (PHVs) with data on patient outcomes with follow-up times of 15 to 20 years or longer is available. The “next generation” of PHVs have some interesting features, but there are no data on patient outcomes at ≥15 to 20 years. The history of PHVs is that: 1) major advances have come in small increments, and 2) extrapolations made from early results were not correct at long term, and when this occurred, patients paid the price in terms of mortality and morbidity. Thus, great enthusiasm from early results and premature prediction may be inappropriate. The data on long-term outcomes are needed and in 2003 one should preferentially select a PHV with proven long-term results.
“Prediction is very difficult, especially about the future.”
—Niels Bohr, Danish physicist, 1885 to 1962
In this issue of the Journal, Bach has presented an interesting article (1)(hereafter referred to as Viewpoint) regarding the choice of the next generation of prosthetic heart valves (PHVs). His hypotheses are that: 1) the recommendations in a recent “review” (2)are inappropriate for the current time because they are based “heavily on the randomized trials” initiated in the late 1970s; and 2) outcomes with use of “newer” PHVs have been shown to be superior, for example, with use of stentless porcine versus stented porcine PHV. Presumably, these hypotheses are meant to be provocative; nevertheless, they need to be examined critically.
The review (2)devoted only about ≤5% of space to the trials. Actually, the review (2)stated the choice of a PHV should be based on several factors, including known long-term results from randomized trials and databases, patient characteristics, expected survival of the patient, and unique patient needs, and also emphasized in the text, Figure 8, and the abstract that in individual patients there may be exceptions to the general rules.
The Viewpoint (1)also stated that in the Department of Veterans Administration (DVA) trial, all the valve-related morbidity with bioprostheses “comprised predominantly of late structural failure.” Actually, at 15 years, after aortic valve replacement (AVR), the incidence of “primary valve failure” comprised only 35% of all complications (3). In addition, the Viewpoint (1)stated that the end point of all valve-related “morbidity” with bioprostheses was superior to mechanical valves “prior to 12 years after surgery” (1). Figure 1 from the Viewpoint (1), which is from the DVA trial, actually shows that for AVR, the total complication rates were virtually identical in the first 2 years and from years 9 to 16. The major cause of the difference in morbidity in the years 3 through 8 was increased bleeding with mechanical PHV, and it was strongly emphasized and discussed in the DVA trial reports that the increased bleeding was at least partly due to the level of anticoagulation, which was excessively high (2,3). Only AVR is emphasized because the Viewpoint mainly discusses this PHV with regard to newer bioprostheses.
With regard to superiority of newer PHVs, space limitations allow a critical evaluation only of the statement that some newer bioprostheses have “statistically significant survival advantages.” The Viewpoint (1)cites three studies (4–6)that have compared mortality at two to eight years after implantation of a stentless versus a stented aortic porcine PHV. The data from the study with the longest follow-up (4)are shown in Table 1. In all three studies (4–6), in both subgroups the late mortality was largely due to deaths from other cardiac and non-cardiac causes, which were greater in the group that received the stented valve. Therefore, it is important to know the associated comorbid conditions at baseline and their treatment that could account for the late deaths (7). However, all three studies (4–6)provided no information about smoking, body mass index, previous myocardial infarction, clinical heart failure, lipids, hypertension, and chronic obstructive pulmonary disease, and two studies (4,6)also did not provide information about diabetes, renal failure, atrial fibrillation, New York Heart Association functional class III and/or IV, and coronary artery disease. A more recent study (8)also was cited in the Viewpoint (1). In this study, the patients were selected from the database of the devices manufacturer (8); moreover, it also has problems similar to the other three cited studies (4–6). Thus, it is questionable whether the current data support the superiority of survival with use of a stentless valve. Indeed, Westaby et al. (5)have stated that “Ideally, the question: ‘do stentless valves convey survival benefit?’ should be addressed by a prospective randomized trial.”
In conclusion: 1) the “next generation” of PHVs must undergo rigorous testing to determine the patient outcomes on long-term follow-up (≥15 to 20 years); 2) these PHVs must be documented to confer the same or better outcomes as the “older” good PHVs currently available; and 3) the long-term results are only partly related to the PHV, for example, only 40% to 50% of total mortality is related to the PHV (2,3). Therefore, one should not compare, or at least be extremely cautious about comparing, outcomes with use of different PHVs, or even the same brand of PHV from different studies, unless all the major and important baseline characteristics (cardiac and non-cardiac) and their treatment are identical or at least very similar (2).
Choice of PHV: 2003
Careful and scholarly review of the literature is important and is advised. Long-term data are not available for the newer prostheses; therefore, it was suggested that clinicians “draw conclusions and make recommendations based on incomplete information, extrapolating from limited data, clinical experience, and common sense” (1). Voltaire once stated that “Common sense is not so common.” Prosthetic heart valves were introduced into clinical medicine 43 years ago, and in the early era (1960s and 1970s) patients undergoing PHV were usually sicker and long-term follow-up data were not available. Therefore, extrapolations from early data to possible longer-term results, although speculative, were appropriate at that time. The history of PHVs is that a very large number of PHVs initially had seemed good and logical extrapolations that were made were later found to be incorrect. As a result, only a few PHVs have lasted, and in those patients in whom the PHVs did not last, the patients paid the price in terms of morbidity and mortality; it needs to be re-emphasized the patient and not the physicians is taking the risks associated with use of PHVs (2). The clinical situation is very different now; a selection of good, albeit not perfect, PHVs with data on patient outcomes with follow-up times of 15 to 20 years or longer is available. This information must be considered in selecting a PHV.
The search for an ideal (perfect) PHV must continue. However, history shows that major advances in PHVs have come in small increments (2)and, thus, one should be very cautious about inappropriate enthusiasm and premature prediction of an expected much better long-term result with “newer” PHVs.
At present, it seems appropriate that: 1) a PHV not yet approved for clinical use by the Food and Drug Administration should be inserted as part of an approved research protocol; 2) if a patient's expected life expectancy is >10 years and one is recommending a PHV with favorable follow-up data of only ≤10 years, then one should take care to accurately and precisely so inform the patient who should also be informed about potential risks and benefits with use of such a PHV; it is prudent and important that the information provided to the patient is adequately documented; and 3) if a patient's expected life expectancy is >10 years, then one should preferably choose a PHV with known favorable patient outcomes data of ≥15 to 20 years.
- Received June 23, 2003.
- Revision received July 1, 2003.
- Accepted July 22, 2003.
- American College of Cardiology Foundation
- ↵Bach DS. Choice of prosthetic heart valves: update for the next generation. J Am Coll Cardiol 2003;42:1717–9
- Rahimtoola S.H.
- Hammermeister K.E.,
- Sethi G.K.,
- Henderson W.C.,
- Grover F.L.,
- Oprian C.,
- Rahimtoola S.H.